detectron2-tutorial

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #golang #rust #deep-learning #linux #docker #macos #java #kubernetes #vue

detectron2-tutorial

(★ 12)

A brief introductory tutorial to the Detectron2 library.

detectron2-tutorial 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Detectron2 Tutorial
A brief introductory tutorial to the [Detectron2](https://github.com/facebookresearch/detectron2) library.

For more information, refer to the [blog post](https://roboticseabass.com/2020/11/22/object-detection-and-instance-segmentation-with-detectron2/)

By Sebastian Castro, 2020

---

# Getting Started

First, you will need a Linux or macOS environment set up with Python 3.6 or higher.

* Create a new Python virtual environment, e.g., `python3 -m venv .virtualenvs/detectron2`
* Activate the virtual environment, e.g., `source .virtualenvs/detectron2/bin/activate`
* `pip install wheel`
* `pip install jupyter torch torchvision`

* Install Detectron2 as outlined in [the Detectron2 install guide](https://github.com/facebookresearch/detectron2/blob/master/INSTALL.md). For Torch 1.7 / CUDA 11.0 (which is what was used for developing this tutorial), the command is:
```
python -m pip install detectron2 -f \
  https://dl.fbaipublicfiles.com/detectron2/wheels/cu110/torch1.7/index.html
```

* Set up OpenCV. Detectron2 requires Python 3.6 or higher so you will have to install from source according to this resource: 
https://docs.opencv.org/master/d2/de6/tutorial_py_setup_in_ubuntu.html

---

# Labeling Data

We have used [coco-annotator](https://github.com/jsbroks/coco-annotator) to manually label our data and export it to the popular COCO data format.

This repository includes a small dataset in the `data` folder so you can try Detectron2 without having to collect and label your own data.

---

# Using Detectron2 for Object Detection

Once your environment is set up, you can follow along the notebooks in this repository.

* [`detectron_pretrained.ipynb`](detectron2_pretrained.ipynb) shows how to use pretrained models from the Detectron2 Model Zoo.
* [`detectron_tutorial.ipynb`](detectron2_tutorial.ipynb) shows how to train a model on a custom dataset by starting from one of the pretrained models above.

detectron2-tutorial

detectron2-tutorial 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

detectron2-tutorial

detectron2-tutorial 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README